Review of possible information platforms for CIRCASA’s Knowledge Information System

ISRIC Report 2018/02

Niels H. Batjes

© 2018, ISRIC – World Soil Information, Wageningen, Netherlands

All rights reserved. Reproduction and dissemination are permitted without any prior written approval, provided however that the source is fully acknowledged. ISRIC requests that a copy, or a bibliographical reference thereto, of any document, product, report or publication, incorporating any information obtained from the current publication is forwarded to:

Director, ISRIC - World Soil Information Droevendaalsesteeg 3 (building 101) 6708 PB Wageningen The Netherlands E-mail: soils@isric.org

The designations employed and the presentation of material in this information product do not imply the expression of any opinion whatsoever on the part of ISRIC concerning the legal status of any country, territory, city or area or of is authorities, or concerning the delimitation of its frontiers or boundaries.

Despite the fact that this publication is created with utmost care, the author(s) and/or publisher(s) and/or ISRIC cannot be held liable for any damage caused by the use of this publication or any content therein in whatever form, whether or not caused by possible errors or faults nor for any consequences thereof.

Additional information on ISRIC can be accessed through http://www.isric.org

Citation

Batjes N.H., 2018. Review of possible information platforms for CIRCASA’s Knowledge Information System. Report 2018/02, ISRIC – World

Soil Information, Wageningen (doi: 10.17027/isric-7Y7B-6S67)

Review of possible information platforms for CIRCASA’s Knowledge

Information System

Niels H. Batjes

ISRIC Report 2018/02

Wageningen, October 2018

Contents

Preface .............................................................................................................................................................................. 5

1 Introduction ............................................................................................................................................................. 7

2 Requirements .......................................................................................................................................................... 8

3 Review of possible data platforms .......................................................................................................................... 9

3.1 Dataverse ........................................................................................................................................................ 9

3.2 PANGEA ......................................................................................................................................................... 12

2.3 GeoNetwork .................................................................................................................................................. 14

2.4 GeoNode ....................................................................................................................................................... 15

4 Ontologies ............................................................................................................................................................. 16

5 Concluding remarks and next steps ....................................................................................................................... 18

Acknowledgements ........................................................................................................................................................ 19

References ...................................................................................................................................................................... 19

List of figures

Figure 1. Example of a Dataverse (INRA, France). .......................................................................................................... 10 Figure 2. Key metadata fields of a Dataverse (https://dataverse.org/) .......................................................................... 11 Figure 3. Main features of a Dataverse. .......................................................................................................................... 11 Figure 4. Example of customised GeoNode instance, ICRAF Landscape Portal. ............................................................. 16 Figure 5. A consistent metadata standard and ontology is needed to exchange data from multiple sources28. ........... 17

ISRIC Report 2018/02 5

Preface

Many studies have shown that agricultural soils have potential to sequester carbon, when judiciously managed, resulting in various initiatives to sequester organic carbon in agricultural lands. Yet, there is still much uncertainty about this potential. The EU H2020 CIRCASA project aims to address this knowledge gap.

The CIRCASA project is implemented by a wide range of partners and coordinated by INRA (Institut national de la recherche agronomique, France). ISRIC - World Soil Information is tasked with the development of a Knowledge Information System (KIS) that will host knowledge on carbon sequestration in agricultural soils. This KIS will include metadata and data from experiments, as well as models and methodological guidelines developed by CIRCASA. ISRIC will also take part in the development of an On-line Collaborative Platform (OCP), a networking tool aiming at bringing together researchers, stakeholders and practitioners in the field. Through these activities, ISRIC will support the improved exchange and accessibility of information on carbon sequestration in agricultural soils.

As a first deliverable, ISRIC prepared a technical report specifying key requirements for developing a KIS for CIRCASA; in this approach, a new comprehensive platform would essentially have to be developed. Pragmatically, however, there are already several operational platforms that could be used as a basis for the knowledge information system. This report aims to provide a brief review of such platforms to inform the decision process.

ir. Rik van Den Bosch Director, ISRIC – World Soil Information

ISRIC Report 2018/02 6

ISRIC Report 2018/02 7

1 Introduction

Awareness that agricultural soils have potential to sequester carbon has resulted in various initiatives to sequester

organic carbon in agricultural lands (Soussana et al. 2017; UNCCD 2017; UNEP 2012). Yet, there is ongoing dialogue

on the potential for sequestration in agricultural lands and a need for sharing knowledge and experiences on how to

make this happen (FAO and ITPS 2015; Harden et al. 2017; Sulman et al. 2018). The EU H2020 CIRCASA project (2017-

2020) aims to address this knowledge gap.

The overall objective of CIRCASA is to strengthen synergies among researchers and promote the transfer of

knowledge on carbon sequestration in agricultural soils. As indicated on the CIRCASA1 website, this will be achieved

through four complementary activities that may be summarised as follows: a) strengthen the international research

community, b) improve our understanding, c) co-design a strategic agenda, and d) create an international research

consortium. The project, coordinated by INRA, is implemented by 24 partners.

Through the above activities CIRCASA will contribute to the implementation of the 2030 Agenda for Sustainable

Development and the Paris agreement within the UN Framework Convention on Climate Change (UNFCCC). CIRCASA

will achieve these goals in synergy with several international initiatives, such as The Global Alliance on Agricultural

Greenhouse Gasses (GRA), the Joint Programming Initiative on Sustainable Agriculture, Food Security and Climate

Change (FACCE-JPI), and the four per 1000 Soils for Food Security and Climate Initiative (4p1000). Further, CIRCASA

will benefit from the CGIAR research programmes on Climate Change Agriculture and Food Security (CCAFS) and

Water, Lands and Ecosystems (WLE).

The CIRCASA project will better structure current knowledge on agricultural soil carbon by:

i) Knowledge synthesis activities combining data (e.g. meta-analysis of scientific literature), integrating

modelling results, developing methodological guidelines (e.g. on monitoring, reporting and verifying

changes in agricultural soil carbon stocks),

ii) Co-designing with stakeholders a pilot knowledge system on agricultural SOC sequestration integrated into

the Online Collaborative Platform (OCP). This integrated system will include geo-referenced meta-data and

(when possible) data from experiments, observations and surveys, as well as from models and synthesis

activities and methodological guidelines developed by CIRCASA. Ultimately, these tasks will lead to an

enhanced international knowledge system (KIS) delivering improved scientific resources of both global and

local significance (e.g. maps showing the technological potential for SOC sequestration of diverse

agricultural practices).

Ultimately, the goal is to enable seamless integration of the discovered data with models and analytical tools. Users should be able to find easily resources via the OCP and to explore them visually. A prototype KIS will be set up for this purpose; it will provide access to geo-referenced meta data describing data from experiments, observations, surveys, crowd-sourcing, as well as from models and knowledge synthesis.

1 https://www.circasa-project.eu/

ISRIC Report 2018/02 8

The requirements for the KIS are described in CIRCASA deliverable D1.2 (CIRCASA/ISRIC 2018); this is a solid yet

rather technical and conceptual document. The elaborate system would be delivered by ISRIC in month 24 of the

project. However, following on discussions with the CIRCASA secretariat, ISRIC was asked to undertake a short review

of freely available platforms for hosting metadata that may later be queried from the OCP. By their nature, these

platforms may be less comprehensive than specified in the above ‘requirements report’.

This report consists of four sections. Main requirements for the KIS as inventoried during an earlier study

(CIRCASA/ISRIC 2018), hereafter referred to as DocD1.2, are summarised in Section 2. Subsequently, in Section 3,

four potential open-source systems (metadata catalogues) for storing metadata in an inter-operable way, are

discussed, pointing at their respective merits and possible limitations in relation to the overall CIRCASA objective. The

need for a consistent ontology, and possible options for this, is discussed in Section 4. Concluding remarks and follow

up actions are formulated in Section 5.

2 Requirements

Requirements of records to be considered in the KIS are detailed in Appendix A of DocD1.2, as prepared by Luis de Sousa (ISRIC) based on consultations with the CIRCASA consortium. These include: name (description of the requirement), priority (degree of importance, following the MoSCow scale: Must Have, Should Have, Could Have, Won't Have this time); type; subject (user domain, user profiles, processes); life phase; risk level (risk of failing to meet the requirement); rationale (for the requirement, textual); verifiable (a Boolean indicating whether the requirement is verifiable or not); and verification (indicating how the requirement can be verified).

The system will focus primarily on scientific information to support the scientific process and the development,

management and evaluation of SOC policy, while other knowledge assets, that may be easier to understand by the

general public, will be shared among network members using the emerging information sharing tools of the OCP. By

its nature, the KIS is not aimed at storing data, but rather at referencing and facilitating on-line access to the various

resources, as held in various platforms (see Section 3), using automatic search/download through queryable

metadata. DocD1.2 requires adoption of OGC (Open Geospatial Consortium) web service standards (WMS, WCS and

WFS) in this regard.

Examples of ‘collaborative platforms’ in the agricultural/environmental domain include those being implemented for the ‘4p1000’ initiative2 and by the 15 CGIAR research centres to ‘mobilize the vast amounts of agricultural research data at their centres and elsewhere to produce new insights and increase the impact of agricultural research for development.’3

Another indispensable element of a dataset is its License as it determines whether a dataset is publicly available or

not, which sets the conditions for its use in scientific studies or policy development.

2 https://hub.4p1000.org/ 3 https://foodtank-com.cdn.ampproject.org/c/s/foodtank.com/news/2018/08/cgiar-big-data-platform-medha-devare/amp/

ISRIC Report 2018/02 9

The number and types of knowledge assets referred by the central KIS is likely to become large, requiring a defined

ontology (see section 3.2 in DocD1.2). By its nature, such an ontology will be dynamic; expert users will be expected

to expand the pool of keywords through a supervised process as described in DocD1.2 (3.2). However, as the

development and testing of a full-fledged ontology is a huge task, CIRCASA may wish to consider adopting an

operational multi-lingual system (see Section 4).

A key process for any KIS is the submission of new datasets or publications. Data providers must first check whether

the material they wish to submit is already accessible on-line. In the affirmative, a reference to its location (e.g. DOI)

or to a service providing access (i.e. URL, ftp address or WFS connection string), will suffice to submit the dataset.

Alternatively, data providers must make sure their dataset complies with the rules set by the Technical Platform

Committee (TCP, see DocD1.2). These concern characteristics aspects such as spatial resolution, coordinate reference

system, and data size/volume. Thereafter, the user has to define the meta-data. A minimum set of meta-data fields is

desired in the meta-data record, and these may vary with the standard adopted for various existing platforms, as

discussed in Section 3. As indicated, the metadata should clearly specify the licence according to which the

data/information is shared (i.e. restricted to open access) and cited.

As indicated, the KIS is one of various deliverables of CIRCASA and it must be properly aligned with the OCP, the

online platform that will function as the doorway to the CIRCASA knowledge base. The KIS must be integrated with

the OCP in a transparent and seamless way, with users largely unaware of the transition between the two, even

unaware of the existence of different technical platforms.

3 Review of possible data platforms

Four freely accessible on-line platforms are reviewed in this section in terms of their suitability for CIRCASA. To

guarantee the long-term persistence of the ultimate system, such platforms should be based on open source

technologies and be themselves open source. Thereby, future development and maintenance of the system will

remain possible once the CIRCASA project ends in 2020. Dedicated communities will then be able to maintain and

further develop the platform thus ensuring that the knowledge collected during the project remains freely available

to the international community.

Metadata for peer-reviewed publications can be harvested from various sources through their DOI’s, based on

selected keywords. These platforms, such as Web of Science, are not discussed here.

3.1 Dataverse

Dataverse4 is an open source web application to share, preserve, cite, explore, and analyse research data. It

facilitates making data available to others, and allows users to replicate others' work more easily. Researchers,

4 https://dataverse.org/about

ISRIC Report 2018/02 10

journals, data authors, publishers, data distributors, and affiliated institutions all receive credit and web visibility

through a Dataverse.

A Dataverse repository is the software installation, which may then host multiple virtual archives called dataverses.

Each dataverse contains datasets, and each dataset contains descriptive metadata and data files (including

documentation and code that accompany the data). As an organizing method, dataverses may also contain other

dataverses.

Dataverse has grown considerably over time and is now a major international collaborative project. At the moment,

there are 34 installations worldwide, corresponding with some 2872 Dataverses holding over 52,000 datasets from

various domains. Some of them are ‘restrictive’ in the sense that they only accommodate data from a given

institution, or in collaboration with a given institution, as illustrated in Figure 1. A demo version is available for

testing purposes5.

According to Dataverse6, ‘any type of quantitative or qualitative data (in any format)’ can be uploaded. However, in

practice, geospatial data may only be submitted as ESRI® shapefiles, thus excluding a wide range of other geospatial

data (polygon and grid based).

By default, Dataverse uses the Creative Commons CC07 waiver for all submitted datasets (4.0 and on). However,

importantly, data depositors can opt-out of using the CC0 waiver for their datasets, if needed, making full use of the

available CC options8 (e.g. CC-BY, by attribution). Metatadata requirements, however, remain the same, irrespective

of the licence.

Figure 1. Example of a Dataverse (INRA, France).

5 https://demo.dataverse.org/ 6 http://dataverse.harvard.edu 7 https://creativecommons.org/share-your-work/public-domain/cc0/ 8 https://creativecommons.org/licenses/

ISRIC Report 2018/02 11

Various integrations are available through which the functionality of Dataverse is gradually expanded, for example

linkage to Matomo/PIWIK9 and ORCID10, see elsewhere11. Geospatial files may be mapped through integration with

WorldMap12; as indicated, this for visualisation of shapefiles only. WorldMap, however, is itself a data harvesting

system, sending all datasets to a server in Harvard; this could conflict with dataset licences.

A dataset in Dataverse is a container for data, documentation, code, and the metadata describing this dataset (Figure

2); main features are summarised in a screen dump (Figure 3). Information on functionality13 and a developer’s

guide14 are available online.

Figure 2. Key metadata fields of a Dataverse (https://dataverse.org/)

Figure 3. Main features of a Dataverse.

9 https://en.wikipedia.org/wiki/Matomo_(software) 10 https://orcid.org/ 11 https://dataverse.org/integrations 12 http://worldmap.harvard.edu 13 http://guides.dataverse.org/en/4.9.2/user/dataset-management.html#file-handling-uploading 14 http://guides.dataverse.org/en/4.9.2/developers/index.html

ISRIC Report 2018/02 12

Overall, Dataverse appears to be a well-supported and user-friendly system for managing a wide range of data,

except geo-spatial data. As such, it is widely used worldwide. A possible shortcoming for CIRCASA, however, is that

Dataverses cannot (yet) handle grid data, such as SoilGrids250m-derived SOC maps, but this may be remedied by

using other (complementary) open source platforms such as GeoNetwork and Geonode. Further, Dataverse has its

own Metadata format15 whereas ISO or OGC standards may be preferred as indicated in DocD1.2.

If selected as a suitable candidate platform for CIRCASA, a project specific (customised) Dataverse may be considered

for partners to add their data respectively to harvest project-relevant data from other Dataverses. Pragmatically, the

central platform should be multilingual to facilitate international collaboration, rather than say mainly in French16 as

is the case for the INRA Dataverse; alternatively, metadata are provided using either French or English, which may be

provide a source of confusion (see Section 4). Further, a drawback of Dataverses is that they are not (yet) OGC

compliant, a key requirement listed in DocD1.2 (CIRCASA/ISRIC 2018).

Dataverses can be searched by broad subjects only, for example ‘soils and soil sciences’, ‘water resources’, ‘farming

systems and practices’ or ‘climate’. Further ‘keyword terms’ can be entered, but no specific thesaurus or ontology is

mentioned. Such would be needed for more in-depth querying and data harvesting (i.e. inter-operability), as

described by Madalli (2015).

An important feature of any data platform for CIRCASA is that the metadata it holds can be queried from the OCP.

This involves as a process of exchanging metadata with other repositories. As a harvesting client, a given Dataverse

can gather metadata records from remote sources. These can be other Dataverse instances or other archives that

support OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting)17, the standard harvesting protocol.

Harvested metadata records will be indexed and made searchable by other users. Clicking on a harvested dataset in

the search results will take the user to the original repository. Inherently, harvested datasets can only be edited in

the source Dataverse installation, in casu by the original (authenticated) data providers. In cases, these may have to

update their keywords to fit with the agreed upon ontology.

Special attention will need to be paid to user management in the selected KIS and its role in the integration with the

OCP.

3.2 PANGEA

PANGAEA® is a data publisher, the services of which are generally open for archiving, publishing and re-usage of data.

The World Data Center PANGAEA18 is a regular member of the ICSU World Data System. As such, it is essentially

different in scope from the distributed Dataverse system.

15 http://guides.dataverse.org/en/latest/user/dataset-management.html#supported-metadata 16 https://www6.inra.fr/datapartage/Gerer/Stocker-les-donnees/Portail-Data-Inra 17 https://www.openarchives.org/pmh/ 18 https://www.pangaea.de/submit/

ISRIC Report 2018/02 13

As indicated on the PANGEA website, any data from earth and life sciences are accepted for archiving. When starting

the data submission process, data providers are redirected to the PANGAEA issue tracker that will assist them in

providing metadata and uploading data files. Communication with PANGEA editors go through the issue tracker.

PANGAEA is archiving and publishing data according to the Open Access Model. Basic operation is covered through

public funding, but PANGEA has to seek for additional funds, in particular for preparing and archiving new data. In

case data are submitted through a project which has publication costs as part of the funding, PANGAEA would

appreciate a small financial contribution for a data submission; other forms of funded collaborations can be

negotiated.

Most of the data managed in PANGEA are freely available and can be used under the terms of the CC license

mentioned on the data set description. A few password protected data sets are under moratorium from ongoing

projects. The description of each data set is always visible and includes the principle investigator who can be

contacted for access to the data.

PANGAEA is an archive for any kind of data from earth system research and thus has no special format requirements

for submissions. Data may be submitted in the authors format and will be converted to the final import and

publication format by the PANGAEA editors. Data providers are requested to keep the following points in mind to

minimize the preparatory work prior to upload:

• For samples taken or measurements made somewhere on earth, the provision of position(s) is

mandatory (latitude/longitude in decimal degree is preferred).

• If data are supplementary to a publication, the (preliminary) citation with journal title and abstract must be

added.

• Submit ONE issue per publication supplement; several files can be attached to one issue (max. size per file = 100

MB)

• If data are related to a project (where PANGAEA is the designated archive) add the project acronym as label.

• Date/Time must be provided in ISO-format (e.g. 1954-04-07T13:34:11).

• Parameters are always accompanied by a unit.

• Abbreviations should be explained.

• Extended documentations may be added as plain text or pdf-file.

• Submit data tables as excel or tab-delimited text files; specific formats (e.g. shape, netCDF, segy ...) may be

added in zip-archive. It is not apparent if large grid data sets can be processed.

• Submit via the PANGAEA ticket system.

As indicated, all data and metadata are quality checked, harmonized, and processed for machine readability by

PANGEA editors, unlike for submissions to Dataverses.

ISRIC Report 2018/02 14

PANGAEA offers a wide range of web services (SOAP / REST). This includes OAI-PMH for metadata harvesting, which

is an important feature if metadata are to be accessed by the OCP. The API allows to retrieve any set of numerical

and textual data. All PANGAEA datasets also ship with Schema.org / Dataset metadata.

It should be noted that PANGAEA supplies metadata to the data portal of the ICSU World Data System and takes care

of the WDS search engine (ICSU-WDS).

Visualisation of georeferenced data in PANGEA through GIS (Geographical Information System) functionality is

enabled by using Google Earth. Whether this functionality meets the specific requirements for CIRCASA’s KIS

(DocD1.2), remains to be assessed by an IT expert.

During data entry keywords can be assigned to a data submission. Apparently, there is no prescribed thesaurus for

this in PANGEA. Further, ‘keywords can be freely defined by the curator’19.

Licences for all submissions are to be provided according to Creative Commons (version 4.0 International) except for

CC0 (1.0 Universal).

In how far data managed in PANGEA can easily be integrated with the OCP remains to be assessed, based on a

technical review.

2.3 GeoNetwork

GeoNetwork is a catalogue application to manage spatially referenced resources, vector and grid. It provides

powerful metadata editing and search functions as well as an interactive web map viewer. It is currently used in

numerous Spatial Data Infrastructure (SDI) initiatives across the world, for example ISRIC’s soil datahub20 or FAO’s

Geonetwork21. Entry of new metadata, however, will require a login (which may be restricted to specific

communities).

GeoNetwork provides an easy to use web interface to search geospatial data across multiple catalogues. The search

functionality provides full-text search as well as faceted search on keywords, resource types, organizations, scale, etc.

Users can easily refine the search and quickly get to the records of interest.

GeoSpatial layers, but also services, maps or even non-geographic datasets can be described in the catalogue. Users

can easily navigate across records and find sources or services publishing a dataset. Metadata on reports or papers

that describe the methodology/study can be presented, with online access to scanned sources (e.g. PDFs or via DOI).

The interactive map viewer of GeoNetwork is based on OpenLayers 3. It is able to access to OGC22 services (WMS,

WMTS) and standards (KML, OWS). Connected to the catalogue, users can easily find new services, layers and even

19 https://wiki.pangaea.de/wiki/Thesaurus 20 http://data.isric.org/geonetwork/ 21 http://www.fao.org/geonetwork/srv/en/main.home 22 http://www.opengeospatial.org

ISRIC Report 2018/02 15

dynamic maps to combine them together; this functionality is not provided by Dataverse or PANGEA. User maps can

be annotated, printed, and shared with others.

The Geonetwork editor also provides multilingual metadata editing, a validation system (e.g. against INSPIRE23

requirements), suggestions to improve metadata quality, geo-publication of layers to publish geodata layers in OGC

services (e.g. GeoServer).

GeoNetwork allows harvesting24 from many sources including: OGC-CSW 2.0.2 ISO Profile, OAI-PMH, Z39.50

protocols, Thredds, Webdav, Web Accessible Folders, ESRI GeoPortal, Other GeoNetwork nodes. These should permit

flexible communication with CIRCASA’s OCP.

According to Lauregui25, a RDF thesaurus can be uploaded in GeoNetwork; subsequently the corresponding keywords

can be used to tag the metadata. These tags can then be used as search criteria. However, GeoNetwork does not use

any of the more interesting semantic information from the thesaurus (e.g., it does not consider information about

subclasses, synonyms, and other thesaurus relations). New Data Catalogue Vocabulary services under development

at e.g. oSGEO26 should be assessed on their merit for handling ontologies by a specialist or CIRCASA’s TCP.

In how far data managed in PANGEA can be integrated with the OCP remains to be assessed, based on a technical

review like for the other platforms under review.

2.4 GeoNode

GeoNode27 is a web-based geospatial content management system. By combining information found in social

networks with specialized geospatial tools, GeoNode makes it easy to explore, process, style, and share maps and

geospatial data.

Spatial datasets can be imported and shared, all through a non-technical user interface. Features include powerful

spatial search engine, federated OGC services, and metadata catalogue.

Geonodes are designed to be extended and modified, and can be integrated/harvested into existing platforms.

GeoNode makes it easy to upload and manage geospatial data on the web. Any user can upload and make content

available via standard OGC protocols such as Web Map Service (WMS) and Web Feature Service (WFS). Data is

available for browsing, searching, styling, and processing to generate maps that can be shared publicly or restricted

to specific users only. Figure 4 gives an example of a customised GeoNode instance as developed for the ICRAF

Landscape Portal.

23 https://inspire.ec.europa.eu/requirements-inspire-directive/64 24 https://geonetwork-opensource.org/manuals/2.10.4/eng/users/managing_metadata/harvesting/index.html 25 https://groups.google.com/forum/#!topic/geonode-users/AhjhEArAbZ8 26 https://trac.osgeo.org/geonetwork/wiki/proposals/DCATandRDFServices 27 http://geonode.org/

ISRIC Report 2018/02 16

Figure 4. Example of customised GeoNode instance, ICRAF Landscape Portal.

Supported upload formats include shapefile, GeoTIFF, KML and CSV. In addition, it is possible to connect to existing

external spatial databases and services.

Features of a Geonode include: publish raster, vector, and tabular data; manage metadata and associated

documents; securely or publicly share data; versioned geospatial data editor.

GeoNode comes with helpful cartography tools for styling and composing maps graphically. These tools make it easy

for anyone to assemble a web-based mapping application with functionality traditionally found in desktop GIS

applications.

Users can gain enhanced interactivity with GIS-specific tools such as querying and measuring. GeoNode’s

documentation is provided at http://docs.geonode.org/en/master/.

GeoNode provides a JSON API which currently supports the GET method. The API is also used as main search engine.

Like for the other platforms under review, a technical review is needed to assess in how far data managed in

GeoNode can be integrated with the OCP.

4 Ontologies

Although not referring to any specific platform for handling metadata, some observations are made here concerning

the need for a consistent ontology as without it user management requirements cannot be met (CIRCASA/ISRIC

2018). Ontology refers to a ‘set of concepts and categories in a subject area or domain that shows their properties

and the relations between them’.

ISRIC Report 2018/02 17

In the case of CIRCASA, data in various formats will be sourced from different organizations using different

information technology facilities, posing a serious challenge as to how to integrate the data into a workable system

so that diverse user groups can query the system unambiguously. One of the most important requirements of such

an integration is that the data semantics are consistent among the different phases of the process. For this, a

consistent ontology should be utilized by all the information systems of the different phases.

The need for a ‘solid, unambiguous’ ontology has been illustrated in a test case for a GeoNode for Malawi28. In this

example it appeared that the ontology of the site did not match users’ mental models for the data’s ontology

(subject categories that show their properties and inter-category relationships).

Every discipline creates ontologies to limit complexity and organize information into data and knowledge. As new

ontologies are made, their use hopefully improves problem solving within that domain (Bonacin et al. 2016; Hu et al.

2011; Su et al. 2012; Zheng et al. 2012). Accessing research papers within every field is made easier when experts

from different countries/institutions maintain a controlled vocabulary of jargon between each of their languages. In

view of the above, choice of an agreed ontology for CIRCASA is considered critical irrespective of the platforms

selected, this in accord with decisions of the TCP. Preferably, the adopted ontology should be compatible (i.e.

interoperable) with procedures adopted for say the 4p1000 OCP or the CGIAR big data platform29.

INRA adopted the semantic web’s practices and standards (RDF, SKOS, OWL, SPARQL) to enable the methodological and technical practices needed by INRA's scientists to standardize, document and publish the vocabularies created in their projects (Jonquet et al. 2018).

Figure 5. A consistent metadata standard and ontology is needed to exchange data from multiple sources28.

28 http://geonode.org/docs/GeoNode_Homepage_UabilityAnalysisReport_March2017.pdf 29 https://bigdata.cgiar.org/ontologies/

ISRIC Report 2018/02 18

Collaborative efforts include AgroPortal, a vocabulary and ontology repository for agronomy, food, plant sciences and biodiversity (Jonquet et al. 2018). This platform already hosts 64 ontologies from varied areas related to agriculture, including AGROVOC30 and the AnaEE thesaurus of INRA. According to Jonquet et al. (2018), the AgroPortal specifically satisfies requirements of the agronomy community in terms of ontology formats (e.g., SKOS vocabularies and trait dictionaries) and supported features (offering detailed metadata and advanced annotation capabilities). Through AgroPortal31, ontology-based projects can be browsed/searched, new projects created, and different ontologies mapped. As illustrated in Figure 5, a consistent metadata standard and ontology are needed to effectively exchange data from multiple sources28.

5 Concluding remarks and next steps

Based on this short, non-technical, evaluation it appears that none of the platforms under review fulfil the

comprehensive specification of requirements for the KIS as presented in DocD1.2 (CIRCASA/ISRIC 2018). In particular,

some of the ‘must have’ requirements on the MoSCoW scale, such as OGC compliance, are not met by some

platforms.

The KIS for CIRCASA will have to be able to handle diverse data types (publications and documents, point data,

polygon data, grid data), use an unambiguous search system, have defined user and user profiles/roles, provide

dataset access within user profiles, and follow consistent standards to permit seamless integration with the

upcoming OCP. Developing a comprehensive, ‘state-of-the-art’ system meeting at least all ‘must have’ and ‘should

have’ requirements would require quite some software-engineering which, realistically, may be beyond ISRIC’s niche.

A customised Dataverse implementation may seem a pragmatic solution for CIRCASA’s knowledge information

system, complemented with GeoNetwork as a versatile, metadata catalogue for handling a broader range of spatial

data. Being a collaborative H2020 project, the platform(s) should be customised with the CIRCASA logo (i.e. not any

particular organisation), with the understanding though that the Dataverse will need to be sustained by a host

institution after completion of the project.

A perceived advantage of these two platforms is that they are open-source, hence maintained by a wide and dynamic

community. However, only GeoNetwork follows OGC standards as recommended in DocD1.2.

For ease of use, preference may be for platforms that use English as the recommended language or have multi-

lingual capability.

CIRCASA partners should make use of, or build upon, an existing, multi-lingual ontology such as maintained by AgroPortal, in accord with TCP’s decision on the matter.

Use of Creative Commons licences is advocated for the various data sources; this is in line with the recommendations of the ICSU World Data System.

30 http://aims.fao.org/vest-registry/vocabularies/agrovoc 31 http://agroportal.lirmm.fr/

ISRIC Report 2018/02 19

A strategic choice has to be made by the CIRCASA management team on how to proceed in relation to deliverable

D1.5 (operational pilot knowledge information system), with a possible re-allocation of tasks, and the integration into

the OCP. A practical solution would be to adopt Dataverse for the KIS and contract a software-engineer to develop an

OGC-compliant integration or plug-in to handle a wider range of spatial data.

Acknowledgements

This document was prepared in the framework of the EU 2020 CIRCASA project. I thank Luis de Sousa for constructive comments, and discussion, on an earlier version of the document.

The findings were discussed during a Skype meeting with representatives of INRA and ISRIC (25 October 2018) to identify the way forward.

References

Bonacin R, Nabuco OF and Pierozzi Junior I 2016. Ontology models of the impacts of agriculture and climate changes on water resources: Scenarios on interoperability and information recovery. Future Generation Computer Systems 54, 423-434. https://doi.org/10.1016/j.future.2015.04.010

CIRCASA/ISRIC 2018. Knowledge Information System Requirements Specification (deliverable D1.2; Compiled by Luís de Sousa/ISRIC), Coordination of International Research Cooperation on Soil Carbon Sequestration in Agriculture (CIRCASA), Wageningen. https://www.circasa-project.eu/content/download/3660/35410/version/1/file/D1.2.pdf

FAO and ITPS 2015. Status of the world's soil resources (SWSR) - Main report, Food and Agriculture Organization of the United Nations and Intergovernmental Technical Panel on Soils, Rome, 650 p. http://www.fao.org/3/a-i5199e.pdf

Harden JW, Hugelius G, Ahlström A, Blankinship JC, Bond-Lamberty B, Lawrence CR, Loisel J, Malhotra A, Jackson RB, Ogle S, Phillips C, Ryals R, Todd-Brown K, Vargas R, Vergara SE, Cotrufo MF, Keiluweit M, Heckman KA, Crow SE, Silver WL, DeLonge M and Nave LE 2017. Networking our science to characterize the state, vulnerabilities, and management opportunities of soil organic matter. Global Change Biology 24, e705-e718. https://doi.org/10.1111/gcb.13896

Hu S, Wang H, She C and Wang J 2011. AgOnt: Ontology for Agriculture Internet of Things. Computer and Computing Technologies in Agriculture IV. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 131-137.

Jonquet C, Toulet A, Arnaud E, Aubin S, Dzalé Yeumo E, Emonet V, Graybeal J, Laporte M-A, Musen MA, Pesce V and Larmande P 2018. AgroPortal: A vocabulary and ontology repository for agronomy. Computers and Electronics in Agriculture 144, 126-143. https://doi.org/10.1016/j.compag.2017.10.012

Madalli DP 2015. Thematic harvesting of agricultural resources from generic repositories. Information Processing in Agriculture 2, 93-100. https://doi.org/10.1016/j.inpa.2015.05.002

Soussana J-F, Lutfalla S, Ehrhardt F, Rosenstock T, Lamanna C, Havlík P, Richards M, Wollenberg E, Chotte J-L, Torquebiau E, Ciais P, Smith P and Lal R 2017. Matching policy and science: Rationale for the ‘4 per 1000 - soils for food security and climate’ initiative. Soil and Tillage Research. https://doi.org/10.1016/j.still.2017.12.002

Su X-l, Li J, Cui Y-p, Meng X-x and Wang Y-q 2012. Review on the Work of Agriculture Ontology Research Group. Journal of Integrative Agriculture 11, 720-730. https://doi.org/10.1016/S2095-3119(12)60061-6

Sulman BN, Moore JAM, Abramoff R, Averill C, Kivlin S, Georgiou K, Sridhar B, Hartman MD, Wang G, Wieder WR, Bradford MA, Luo Y, Mayes MA, Morrison E, Riley WJ, Salazar A, Schimel JP, Tang J and Classen AT 2018. Multiple models and experiments underscore large uncertainty in soil carbon dynamics. Biogeochemistry. https://doi.org/10.1007/s10533-018-0509-z doi:10.1007/s10533-018-0509-z

UNCCD 2017. The Global Land Outlook (First Edition), United Nations Convention to Combat Desertification, Bonn, 336 p. https://knowledge.unccd.int/sites/default/files/2018-06/GLO%20English_Full_Report_rev1.pdf

UNEP 2012. The benefits of soil carbon - managing soils for multiple, economic, societal and environmental benefits, UNEP Yearbook - Emerging issues in our global environment 2012. United Nations Environmental Programme, Nairobi, pp 19-33

Zheng Y-l, He Q-y, Qian P and Li Z 2012. Construction of the Ontology-Based Agricultural Knowledge Management System. Journal of Integrative Agriculture 11, 700-709. https://doi.org/10.1016/S2095-3119(12)60059-8

www.isric.org

Together with our partners we produce, gather, compile and serve quality-assured soil information at global, national and regional levels. We stimulate the use of this information to address global challenges through capacity building, awareness raising and direct cooperation with users and clients.